V Kakulapati*
Sreenidhi Institute of Science and Technology, Yamnampet, Ghatkesar, Hyderabad, Telangana, India
*Corresponding Author: V Kakulapati, Sreenidhi Institute of Science and Technology, Yamnampet, Ghatkesar, Hyderabad, Telangana, India
Received: August 09, 2024; Published: September 05, 2024
Citation: V Kakulapati. “Intelligent Modelling of Agricultural Cart for Precision Nutrient Application - A Fusion of Fuzzy Logic and Machine Learning". Acta Scientific Agriculture 8.10 (2024): 03-10.
The Indian agriculture industry is experiencing growth due to the COVID-19 pandemic, but the importance of organic food presents growth opportunities. Online marketing and precision farming techniques are crucial for optimal crop yields. Harvestify, an online application, provides crop and fertilizer recommendations based on soil information, enhancing agricultural productivity and decision-making for farmers. The problem statement aims to create an intelligent modeling system for exact nutrient administration on agricultural carts, combining fuzzy logic and machine learning methodologies to maximize nutrient spread. The system must dynamically modify fertilizer application rates in real time to reduce waste and environmental impact. This sophisticated decision-making framework will lead to increased productivity and sustainability in contemporary farming methods, adjusting to the intricacies and uncertainties present in agricultural operations.
Keywords: Agronomy; Nutrient; Management; Crop Selection; Soil Analysis; Yield Prediction; Seed Optimization; Agriculture; Technology; Sustainability; Profitability
Agriculture serves as the primary livelihood for approximately 58% of India's population, leveraging its vast land area of over 1.6 million square kilometers. With the potential to emerge as a significant force in agriculture, India relies heavily on this sector, which not only caters to the nation's food needs but also stands as one of the leading exporters of agricultural goods globally [1]. Despite its pivotal role, the agricultural industry has encountered various challenges, including those exacerbated by the COVID-19 pandemic such as restrictions on travel and trade. This global crisis has disrupted food demand and supply chains, prompting farmers to make uninformed choices regarding crop selection and leading to inefficiencies in resource utilization. Various machine learning algorithms suitable for agriculture have distinct advantages and drawbacks [2]. Our platform offers recommendations for optimal crops and fertilizers based on soil data input. Through the Crop Recommendation feature, farmers can obtain predictions for suitable crops based on their soil data. Similarly, the Fertilizer Suggestions module analyzes soil information and crop type to provide insights into soil deficiencies or excesses, along with suggestions for improvement. Accessing accurate crop forecasts enables farmers to make informed decisions regarding crop selection.
There can be several detrimental effects on natural resources, including soil, water, and air when individuals are ignorant about scheduling and using appropriate planting practices. Farmers must employ efficient and sustainable planting strategies because variations in the seasonal environment have the potential to worsen these problems [3]. To help farmers maximize their agricultural yields while limiting their impact on the environment, machine learning algorithms can be quite helpful here.
Through the utilization of the Crop Recommendation tool, farmers can input their soil data to receive tailored suggestions on suitable crops for their specific soil and climate conditions. This assists in avoiding the wastage of time and resources on crops unsuitable for their land. Furthermore, the provision of fertilizer recommendations based on soil data aids farmers in optimizing fertilizer usage and enhancing soil health [4]. By inputting soil data and crop type into the Fertilizer Suggestions tool, farmers can obtain advice on nutrient deficiencies and suitable fertilizers to address them, thereby avoiding over- or under-fertilization. Utilizing soil data for crop and fertilizer recommendations also promotes environmental sustainability by enabling resource optimization and waste reduction [5]. This fosters more eco-friendly farming practices and mitigates pollution and nutrient runoff. Offering crop and fertilizer guidance using soil data enables farmers to make informed choices, enhancing crop yields, reducing environmental harm, and fostering sustainable farming methods.
To establish links between the results of soil tests and the trimming response to fertilizers, as well as to set up a calibration for soil test-based fertilizer recommendations. To obtain various fertilizer tests as a basis for recommending fertilizers for yield-focused crops. to evaluate the viability of various soil test techniques in real-world settings. To evaluate the combination application of synthetic and organic fertilizers for enhanced supplement use efficiency. Using the initial soil test results as a basis, deduce a fertilizer recommendation for a complete pruning group.
The goal of the proposed system is to create an intelligent modeling system that will allow for exact nutrient administration on agricultural carts. The process involves the amalgamation of fuzzy logic and machine learning methodologies to maximize the spread of nutrients, considering multiple parameters like crop kind, soil composition, and environmental circumstances. To maximize crop health and productivity and reduce waste and environmental effects, the system must be designed with the ability to dynamically modify fertilizer application rates in real time. With the help of machine learning and fuzzy logic, a sophisticated framework for making decisions will be developed that can adjust to the intricacies and uncertainties present in agricultural operations [6], ultimately leading to increased productivity and sustainability in contemporary farming methods.
The CART method, which is a decision tree algorithm, has a cross-validation feature that identifies over-fitting and so prevents inaccurate forecasts for the future [7]. This characteristic has resulted in more precise forecasts compared to other statistical methodologies. Nevertheless, comprehending the efficacy of these methods continues to be a difficulty in determining their optimal use.
Land records can be used to classify land. Random Forest is the machine learning method used for soil classification. The outcome of the method will be the separation and presentation of the confusion matrix, precision and mean, and finally precision. Using crop, feeding, and source data, a crop yield forecast can be created [8]. Refer to this tactic as the random forest algorithm. Based on the available input, the algorithm will forecast the crop. Fertilizer data, crop data, and location data can all be used to provide fertilizer recommendations. For every crop, appropriate products and fertilizers are suggested in this section. To access temperature data, humidity, barometric pressure, and all descriptors, third-party apps are utilized.
The remainder of the paper is as follows, Sec 2 provides related works of the proposed methodology, whereas, Sec 3 discusses the methodology of the work. The experimental analysis is discussed in sec 4. Sec 5 provides the concluding remarks followed by the future work possibilities discussed.
The literature survey on the intelligent modeling of agricultural carts for precision nutrient application, combining fuzzy logic and machine learning, reveals a growing interest in leveraging advanced computational techniques to enhance agricultural practices. Studies have explored various aspects of this integration, focusing on improving nutrient management, crop yield, and resource efficiency. Researchers have investigated the application of fuzzy logic to capture the inherent uncertainties and imprecisions in agricultural data [9], allowing for more flexible decision-making processes. Concurrently, machine learning techniques such as neural networks and support vector machines have been utilized to analyze large datasets and extract meaningful patterns to optimize nutrient application strategies.
Additionally, the literature highlights the significance of real-time monitoring and control systems in agricultural carts, enabling adaptive nutrient application based on dynamic environmental conditions and crop requirements. Studies have proposed intelligent modeling frameworks that integrate fuzzy logic and machine learning algorithms to develop predictive models capable of adjusting nutrient application rates on the fly. These models consider diverse factors such as soil characteristics [10], weather patterns, crop growth stages, and historical data to make informed decisions, ultimately leading to improved crop health and yield. Overall, the literature underscores the potential of combining fuzzy logic and machine learning in agricultural cart modeling to revolutionize precision nutrient management and drive sustainable farming practices.
India is a farming nation, but it continues to supply its farmers using age-old techniques. These days, expert advice is derived from farmer-expert exchanges, and the opinions of many experts vary. It is possible to recommend plants and fertilizers using smartphones running the Android operating system, which is the most popular technology. Mobile devices like tablets and phones, as well as other domestic platforms, are the primary users of the Android operating system. A soil's quality is determined by its nitrogen, potassium, and phosphorus (NPK) content. Based on this, fertilizers can be added to the soil through soil testing to increase agricultural yields [11]. You can write an application for the Android operating system that uses the soil NPK test to suggest various fertilizers. The goal of this system is to close the divide between technology-particularly Android technology-and agriculture. The website itself makes it simple to buy the suggested fertilizers. Online retailers offer a variety of recommended fertilizers for purchase.
This research investigates the influence of climate-responsive factors on agricultural output in Bangladesh by using fuzzy logic. The prediction of agricultural productivity relies on several language factors, including temperature, meteorological disasters, water availability, monsoon level, illnesses, species extinction, and deforestation [12]. The findings indicate a negative correlation between rising temperatures weather-related calamities, and a decline in agricultural output. The impacts of temperature and water availability are uniform, with rising temperature counterbalanced by the presence of water. Monsoon levels lead to significant amounts of rainfall, but moderate temperatures and abundant access to clean water contribute to moderate levels of agricultural productivity. Nevertheless, the loss of species has enduring consequences, whereas deforestation has rapid effects.
The research used fuzzy membership functions to classify ripe tomatoes according to their color, size, and firmness. The samples were classified into five groups: "grade I", "grade II", "grade I-far market", "processing", and "storage". A total of 81 fuzzy rules were reduced to 25 by using suitable rules. A total of six distinct algorithms, each using various fuzzifiers and defuzzifiers, were used. The outcomes obtained from these methods were then compared to the classifications provided by the panelists [13]. The findings demonstrated that fuzzy algorithms effectively categorized fruits into the right groups, yielding optimal outcomes when using zmf and sigmf as membership functions, gbellmf as the fuzzifier, and mom as the fuzzified. The research revealed that by using fuzzy membership functions, the combination of tomato attributes may effectively and precisely categorize tomatoes into appropriate groupings for various markets.
Fuzzy Logic may also be used to assess the quality of various crops, including [14] tomatoes [15], lettuce, cauliflower, and mangoes [16]. These methods include taking or entering images, extracting features, and then classifying and/or grading them [17]. The crop is graded based on factors such as size, form (Mustafa et al., 2009), color, fragrance, etc. The grading is done on a scale ranging from 1 to 10. Furthermore, farmers may effectively allocate their resources by grading date trees according to their condition and expected productivity [18].
The system aims to assist farmers in cultivating high-yield crops by selecting varieties commonly grown in specific regions, considering local agricultural practices and climate conditions. Common crops may include Rice, Jowar, Wheat, Soybean, Sunflower, Cotton, Sugarcane, Tobacco, Onion, and Dry Chili, among others, depending on the area. To ensure accuracy, credible sources such as government agencies, research institutions, and historical records provide crop yield data for dataset creation. This dataset includes vital information like crop type, yield, nutrients, and location for soil classification, crop yield prediction, and fertilizer recommendation processes.
Land classification involves utilizing land records, and employing the random forest algorithm to generate a confusion matrix, precision, and mean output.
Classification and regression trees (CART), also known as decision trees, are decision support tools used in agriculture for disease diagnosis, crop monitoring, and weed classification. They use a tree-like graph or decision model to represent test results and class labels, such as categorizing based on water stress, weed presence, and nitrogen application rates. The proposed work cart achieved 85.3% accuracy in predicting the results.
Fuzzy Logic is a computational method that uses algorithms to determine and forecast the best values of N-P-K fertilizers for promoting maximum crop development. This technology assists farmers in achieving higher crop yields while minimizing expenses on fertilizers. The system administers precise quantities of nutrients to the crop, resulting in reduced fertilizer use, prevention of overfertilization, and mitigation of environmental risks. This technique optimizes fertilizer use, decreases agricultural yield, and mitigates soil deterioration, thereby diminishing environmental hazards. Fuzzy Analysis achieved 93.8% in predicting fertilizer recommendation.
Fuzzy logic plays a vital role in agricultural systems by dealing with uncertainties and imprecisions related to elements such as soil composition, crop needs, and environmental circumstances. It facilitates adaptable decision-making processes, enabling informed modifications to nutrient application rates. This integration improves the accuracy and efficiency of nutrient delivery tactics, enabling the agricultural cart to negotiate complicated situations with human intelligence.
Continuous Improvement: Iteratively refine and update the guidance system based on feedback, new research findings, and changing agricultural practices. This continuous improvement process ensures the system remains relevant and effective over time.
Figure 1: User Interface for entering Crop details. Recommendation of the name of the crop that is suitable to grow in the conditions given by the user.
Figure 2: Prediction of Recommendation crop to farmers.
Figure 3: User Interface for Fertilizer details.
Figure 4: Result Page for Fertilizer Recommendation.
Figure 5: User Interface to upload picture for Disease prediction.
Figure 6: The result of Crop disease prediction.
Figure 7: Result Page for recommendation for Disease.
This user-friendly system aims to bridge the gap between agriculture and technology, offering farmers accessible tools for crop selection and management. Through a graphical user interface (GUI), farmers can easily input land details like soil type, weather conditions, and water availability, receiving precise predictions on suitable crops to grow. A key feature is nutrient recommendations, where our system analyzes soil and crop data to suggest the ideal type and number of fertilizers, enhancing nutrient management and crop health for higher yields. Additionally, our system advises on seed quantity for planting, considering factors like planting density and germination rates, thus optimizing seed usage, and cutting costs. It also forecasts yields based on historical and current data, aiding farmers in planning sales strategies for improved financial management. Real-time market price updates for various crops enable informed decisions, potentially leading to better profits by aligning crop selection with market demand. Our system empowers farmers to make informed choices in crop selection, nutrient management, seed usage, and market strategies, contributing to agricultural development and enhancing farmers' livelihoods globally.
In the future, boost crop forecast accuracy by comparing the effects of suitable machine learning techniques and refining a machine learning model using Metaheuristic algorithms, which leads to better performance and fewer mistakes.
Copyright: © 2024 V Kakulapati. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Open Access by
Acta Scientific is licensed under a Creative Commons Attribution 4.0 International License
Based on a work at https://actascientific.com
ff
